ERROR - transformers.tokenization_utils - Using bos_token, but it is not set yet报错

Question

问题遇到的现象和发生背景

尝试调SSAN关系提取模型报错，解决不了

问题相关代码，请勿粘贴截图

all_input_ids = torch.tensor([f.input_ids for f in features], dtype=torch.long)

报错内容
04/14/2022 15:55:16 - ERROR - transformers.tokenization_utils - Using bos_token, but it is not set yet.
04/14/2022 15:55:16 - ERROR - transformers.tokenization_utils - Using eos_token, but it is not set yet.
Traceback (most recent call last):
File "./run_docred.py", line 759, in
main()
File "./run_docred.py", line 699, in main
train_dataset = load_and_cache_examples(args, tokenizer, evaluate=False)
File "./run_docred.py", line 464, in load_and_cache_examples
all_input_ids = torch.tensor([f.input_ids for f in features], dtype=torch.long)
TypeError: an integer is required (got type NoneType)

运行结果及报错内容

我的解答思路和尝试过的方法

我想要达到的结果

Answer 1

bos_token 没有值，在run_docred.py 找一下这个值的定义，看看是怎么得的，好像是读取文本获取的？看看文件里面的内容